Building Towards Self-Driving Codebases with Long-Running, Asynchronous Agents

Aman Sanger, co-founder and CTO at Cursor, will share how Cursor is building long-running coding agents that can autonomously execute more ambitious software tasks. Key Takeaways: Software engineering is quickly shifting to async agents that work independently and report back like colleagues Self-driving codebases will require multi-agent systems that delegate specialized subtasks to the best model for each job Developers will focus on building detailed, verifiable specs that serve as an implementation plan and evaluation suite Industry: All Industries Topic: Agentic AI / Generative AI - Code / Software Generation Technical Level: Technical - Advanced Intended Audience: Data Scientist NVIDIA Technology: Hopper, Blackwell, DGX Cloud #NVIDIAGTC